Will future revision of sb come with flux less vs Tim paste.

[moonbogg]

does no one see the significance of ivy bridge running hotter than sandy bridge even when delidded

There you go with your "facts" again. Hunches and speculation are more valuable when it comes to matters like this.

 

[Puppies04]

You guys don't think it makes more sense that intel used paste to keep the OC performance of the chips limited? Yeah I know I have no friggin PROOF, but I can't get past common sense on this one. With solder instead of paste, Ivy would clock so damn high that everything would be instantly obsolete in it's wake. It wasn't to save money. I think it was to save Sandy Bridge-E.
Who would buy a 3930K when an ivy bridge could hit 5.5 on air easy? Oh, thats right it can't hit 5.5 on air easy because they used F-ing silly putty instead of proper solder.

If you want some "common sense" how about the fact that ib's cores running hotter because of the increased number of transistors packed into a smaller space was predicted way before the chip was even released, it is basic physics. As for the rest of your post regarding 1155 somehow making everything that went for it obsolete I won't even bother dignifying that with an answer.

 

[Borealis7]

not only that, the logical conclusion is that the 22nm Haswell will suffer from the same illnesses.

 

[ShintaiDK]

not only that, the logical conclusion is that the 22nm Haswell will suffer from the same illnesses.

Yes and no.

Intels tick/tock have this sideeffect. SB was designed for 32nm with all its technologies. IB aint designed for 22nm, Haswell is.

So while Haswell will suffer from the same, it will not be as bad.

Not to mention, the dealing with this issue aint new.

 

[Charles Kozierok]

Not to mention, the dealing with this issue aint new.

Nor are the complaints about lack of visible progress in performance upon moving to a new node.

 

[sm625]

If better TIM leads to ivy easily overclocking to 5GHz on air, then the market for SB-E is pretty much ruined. Very few people will want to sacrifice 10% performance for some extra memory bandwidth.

However, if ivy overclocking is limited such as it is, then SB-E performance is relatively flat, which is enough to justify purchase. Considering Intel is a company with a documented history of conspiring to do much worse than this, it is just pure folly and ignorance to rant about kennedy and conspiracy theories. You're talking about a company that has been fined billions for past conspiracies...

Just think of what happens if Intel releases an IVB-E and it turns out to have better TIM and overclocks like a bandit to 5.3GHz on air. How many people do you think will be coming on these forums worshipping Intel and ready to plunk down $200 extra to go with the more expensive platform? I bet quite a few will fall for this. Bottom line it is a simple decision that could generate an extra million or two in profits for intel.

 

[ShintaiDK]

TIM was due to heatcycles. And we already seen several test without IHS. There is no magic hidden.

There is no IB-E. Next step is HW-E.

And mixing conspiracy with antitrust...really? Lets stay to facts rather than tinfoil mentality.

 

[Lonbjerg]

TIM was due to heatcycles. And we already seen several test without IHS. There is no magic hidden.

There is no IB-E. Next step is HW-E.

And mixing conspiracy with antitrust...really? Lets stay to facts rather than tinfoil mentality.

You must be new here...tinfoilhat conspiracies are rampant on this board.

Often used when ones favourites campany looses the performance crown...then the "normal" response on ATF is...well tinfoil posts.

Just look at the "Linus flipped of NVIDIA" thread.

Facts are getting more and more rare inhere...oddly that conincides with AMD's fall from the top

 

[Charles Kozierok]

Not putting out products that cannibalize/obsolete existing markets isn't conspiracy theory stuff. It's sound business.

I also think enthusiasts make a serious mistake in vastly overestimating how important they are to chipmakers. For a while companies like Intel seemed to be actively hostile towards enthusiasts, but they changed and started putting out products that catered to the high-end market. This cycle is now changing again, I think -- not back towards hostility, but more towards apathy. AMD's inability to compete is a big part of it.

 

[Lonbjerg]

Not putting out products that cannibalize/obsolete existing markets isn't conspiracy theory stuff. It's sound business.

I also think enthusiasts make a serious mistake in vastly overestimating how important they are to chipmakers. For a while companies like Intel seemed to be actively hostile towards enthusiasts, but they changed and started putting out products that catered to the high-end market. This cycle is now changing again, I think -- not back towards hostility, but more towards apathy. AMD's inability to compete is a big part of it.

FFS...intel's compete against it self.
If it dosn't make product that make the stockholders happy (high returns)...they will be forced too.

It's sad to see people ponder....with 1/16th of the facts...and then think they can make a "conclusion"---*sigh*

 

[Charles Kozierok]

Sorry, I don't understand what you're saying.

Intel is not going to be maximizing their share price by worrying about whether a handful of people can overclock because of their thermal transfer material.

 

[T_Yamamoto]

Sorry, I don't understand what you're saying.

Intel is not going to be maximizing their share price by worrying about whether a handful of people can overclock because of their thermal transfer material.

This as well.

But that's why they released the K series. For us enthuests. They released the Extreme edition for entuthests

 

[Lonbjerg]

Sorry, I don't understand what you're saying.

Intel is not going to be maximizing their share price by worrying about whether a handful of people can overclock because of their thermal transfer material.

IVB competes with SB.
Hasswell will compete with IVB.

[redacted]

Intel won't slow down...and won't hamper performance.
Why not?
Because their new CPU lines wouldn't sell if they did.
No new sales = unhappy stockowners.

[redacted]

Lon, your poor treatment of other posters is unacceptable. You are hereby on vacation until you can stop taking out your frustration on other posters.
-ViRGE

 

[Magic Carpet]

Intel won't slow down...and won't hamper performance.
Why not?
Because their new CPU lines wouldn't sell if they did.
No new sales = unhappy stockowners.

What is the % of the people, who buy Intel CPUs every year, just because they are faster, you reckon?

 

[moonbogg]

If better TIM leads to ivy easily overclocking to 5GHz on air, then the market for SB-E is pretty much ruined. Very few people will want to sacrifice 10% performance for some extra memory bandwidth.

However, if ivy overclocking is limited such as it is, then SB-E performance is relatively flat, which is enough to justify purchase. Considering Intel is a company with a documented history of conspiring to do much worse than this, it is just pure folly and ignorance to rant about kennedy and conspiracy theories. You're talking about a company that has been fined billions for past conspiracies...

Just think of what happens if Intel releases an IVB-E and it turns out to have better TIM and overclocks like a bandit to 5.3GHz on air. How many people do you think will be coming on these forums worshipping Intel and ready to plunk down $200 extra to go with the more expensive platform? I bet quite a few will fall for this. Bottom line it is a simple decision that could generate an extra million or two in profits for intel.

Yeah, exactly. Better paste cooled the chips quite a bit. If they used real solder they would have been totally killer. It would look ridiculous to have a sub $300 CPU totally smashing the hell out of a 6 month old, $1,000.00 monster of a chip, don't you think? It makes perfect marketing sense that they would limit Ivy in this way. Its not a conspiracy theory. Its an observation that should be easy enough to spot by anyone who cares.

 

[Diogenes2]

Better paste cooled the chips quite a bit.

Evidence ?

 

[Puppies04]

Evidence ?

When you get your "evidence" don't forget to ask if this new TIM will last a minimum of 3 years without reapplying/ degrading performance which the intel stuff has to.

 

[zephyrprime]

But before you go looking keep this basic fact in mind, heat aswell as voltage degrades chips, if intel had to spend a few extra cents on TIM they would because A: it will reduce warranty replacements down the line and B: If IB chips do start degrading because intel save a few thousand bucks on TIM the cost to them could very well be in the millions because of the bad press generated.

This isn't some backyard business trying to cut corners, if they changed the TIM there is a damn good reason for it and as they gave what appears to be (to me anyway) a good reason as soon as they were asked I tend to trust them on this one.

But then why didn't they use a better thermal paste? Also, you are giving them too much credit about trying to cut corners. Time and again, history has shown that NUMEROUS huge companies have shot themselves in the foot and even killed people by cutting very minor corners from their products.

 

[zephyrprime]

does no one see the significance of ivy bridge running hotter than sandy bridge even when delidded

I bet it's because of the fin (tri gate) transistors. They have more exposed surface area between the channel and the gate so it stands to reason that the leakage current would get worse. Maybe they need so more insulator there to rectify the problem or go with a wussier, lower aspect ratio fin.

 

[Diogenes2]

But then why didn't they use a better thermal paste?....

Better than what ?

Where is the evidence there is a better paste available, for the part as designed?

 

[zephyrprime]

When you get your "evidence" don't forget to ask if this new TIM will last a minimum of 3 years without reapplying/ degrading performance which the intel stuff has to.

What makes you think the intel stuff won't degrade in performance? I bet it will.

 

[Diogenes2]

What information do you have about the Intel ' stuff ' ?

 

[Puppies04]

But then why didn't they use a better thermal paste? Also, you are giving them too much credit about trying to cut corners. Time and again, history has shown that NUMEROUS huge companies have shot themselves in the foot and even killed people by cutting very minor corners from their products.

Ok i'm going to try this one more time but I really don't know how I can make it much clearer.

There was no "better" thermal paste they could have used, ok maybe some super exotic stuff that would have added several $ to production costs but lets be realistic here. The reason for the change in TIM is because IB cores get hotter over the smaller cores. Not only have you got increased temperature you also have localised heat with cooler areas around it. Both of those factors meant that Intel could not guarantee that the existing TIM would have held up to 3 years (warranty period) of thermal cycling. People are delidding their chips and still struggling to hit SB temps, i'm not trying to claim there was nothing else intel could have done but when the directors were told that they could reduce temps when overclocking by a further 5 degrees C but it would add lets say $4 to the retail cost of each chip they would have made a call to go with the existing solution.

We (overclockers) aren't and never have been Intels first priority and acting like we are is naive to the extreme.

You can still OC IB quite well, it functions perfectly within spec and is faster and more power efficient than SB with a lot better onboard graphics I call that a win even if the laws of physics are doing their best to make some people call it a failure.

 

[Puppies04]

What makes you think the intel stuff won't degrade in performance? I bet it will.

Maybe the fact every intel chip comes with a 3 year warranty and extremly strenuous tests would have been carried out to make damn sure that in 2 years time 500,000 chips aren't going to have to be replaced for free. Or we can go with you hunch, your choice.

 

[2is]

What makes you think the intel stuff won't degrade in performance? I bet it will.

What makes you think it will?